7/ Survey experiments surge and now account for roughly 45% of all design-based studies. Lab experiments decline sharply after 2016, while field experiments and most quasi-experimental designs grow slowly from a very low base

6/ The revolution is driven mainly by survey experiments and, to a much lesser extent, DID and unconfoundedness designs with matching & reweighting. Field experiments, IV, RDDs, natural experiments, and synthetic control all remain relatively rare in the discipline.

4/ In our coding, design-based studies pin causality on a clear source of (quasi-)random variation and spell out identification assumptions, while model-based studies rely mainly on functional-form and statistical assumptions, often treating regression coefficients as causal

3/ About 40% of articles are empirical & quantitative. Of these, 81% are explanatory–using causal relationships to explain social phenomena–and just 29% of explanatory studies use explicit design-based methods. Most causal work still relies on model-based strategies or leaves the design implicit

🔟 Player-avatar bonds and gaming benefits and risks: Assessing self-discrepancy theory against a broader range of character and play experiences

Snodgrass, Sagstetter, Branstrator, Giardina, Lacy, Bollinger-Deters, Callendar, Zhao, Dengah II, & Billieux

doi.org/10.5817/CP20...